Add MiniMax backbone support to Mobile-Agent-E

[octo-patch]

Reason: add target provider/model to existing provider registry

Summary

• Add MiniMax as a selectable Mobile-Agent-E backbone using the MiniMax-M3 model and OpenAI-compatible endpoint.
• Track MiniMax-M3 usage costs with the configured 0.6/2.4 USD per million token rates.
• Document MiniMax environment variables in English and Chinese READMEs.

Test plan

• python3 -m py_compile Mobile-Agent-E/inference_agent_E.py Mobile-Agent-E/MobileAgentE/api.py
• git diff --check

[octo-patch]

This PR has been replaced by #290 with a clean implementation based on the latest main branch and focused endpoint, protocol, and pricing tests.